Spin Semiconductor

Hi all:

From http://www.spinsemi.com/freeform/INST_SYNTAX.htm#JAM

"This can be used in pitch transposition to intelligently reset the read pointer when the input and output of the transpose function are highly correlated."

2 questions:

- What sort of correlation measure can be performed on the FV-1? I can see a simple zero-crossing detector, but am unsure of how to perform autocorrelation, AMDF or the like.

- Will resetting the RAMP allow for crossfading, or is this assuming a single output pointer?

Thanks,

Sean Costello

I've only ever played with zero-crossing and JAM, would be very interested to see something more complex.

It is a single pointer that is simply reset to the start of the delay so I don't think that it could be used to generate a crossfade coefficient.

How is JAM used? I tried putting the following in some pitch shifter code:

JAM rmp0

If inserted into the main code loop, it seems like this should simply halt any pitch shifting, but it seems to have no effect. Is there some other way of identifying the ramp to be reset?

Thanks,

Sean Costello

Can you post or email complete code so I can see it all in context?

Email w/code sent...

Thanks to Sean for the code and help finding the problem. Issue was an error in SpinAsm not assembling the JAM properly (also documentation is wrong, look for a new SpinAsm doc file in a day or so), a new build is available for d/l.

I think this problem may not be fixed (or I'm confused on how JAM works).

I'm sending RMP0 out the DAC and watching it on a scope. With <wldr 0, 1023, 4096> I get a ramp starting at 2.25V and dropping to 1.5V at a rate of 2 Hz.

I've written a program so JAM is run anytime one of the pots on the dev board is at zero. So with a ramp going, I turn the pot to zero and it seems like the pot has to be held there for 700 ms for the JAM to take place. The waveform goes to 2.25V and stays there until I turn the pot back up, at which point it almost instantly goes back to its normal waveform.

What you describe in terms of operation seems correct. The ramp is forced to the top on a JAM, but the 700mS delay seems wrong. The pot inputs should only have about a 100mS delay.

Please keep in mind that what you see on the output of the DAC will be affected in a few ways, when the ramp is JAM'd it is a static (DC) value, the DACs are delta-sigma converters and have filtering in them.

Also, try a higher frequency coefficient for the ramp, something like:
wldr 0,16384,4096